The present invention concerns the confluence of two issues related to high frequency test equipment, and particular, to test equipment where individual coaxial connectors are used to connect a detachable probe to that equipment. One example is present day high performance oscilloscopes.
The first issue concerns what series connector is used, especially for probes or the connections to signals to be measured by the test equipment, and that are not merely an ancillary part of a test set-up. It is customary for ""scopes to employ BNC connectors for their front and rear panel connections. The BNC connector has a number of attractive features that, so far anyway, have outweighed its disadvantages. These attractive features include ease of use (a quarter twist to mate or un-mate), small enough to not consume too much panel space but not so small as to be mechanically delicate, reasonable in cost and already widely in use with many manufacturers and mounting styles to choose from. It is also a controlled impedance connector, and is available in the commonly used values of 50 xcexa9 and 75 xcexa9. Save for characteristic impedance, any BNC connector will (in theory, anyway) mate with one of the opposite gender, regardless of who the manufacturers were or what the mounting styles are. In many respects it is the workhorse of the general electronics industry; if it wasn""t at hand we""d have to invent it. Nevertheless, and despite its longevity and venerable origin [the Bayonet Navy Connector (BNC) was developed for the US Navy during WW II] it has begun to reveal certain shortcomings. The following several paragraphs relating to the shortcomings of the conventional BNC connector, and an attractive solution therefor, have been abstracted from PRECISION BNC CONNECTOR.
Despite its popularity, the BNC connector has some significant drawbacks when used as an instrument grade connector for some electronic test equipment, such as top of the line high frequency oscilloscopes. It has reactive discontinuities at high frequencies. That is, above certain frequencies it fails to match the 50 xcexa9 characteristic impedance of the coaxial transmission line of which it is expected to be a part. Even the most carefully installed silver-plated mil-spec clamp type BNC connector is extremely visible as a discontinuity on a TDR (Time Domain Reflectometer) of even modest bandwidth. Next, it tends to xe2x80x9cleakxe2x80x9d (radiate from its mating surfaces) above, say, 500 MHZ. Finally, since it relies solely on internally supplied spring tension to draw its parts together, it can, when under externally applied tension, allow the mating parts to separate sufficiently to degrade the quality of the connection (greater discontinuity, more loss), sometimes to point where the connection is interrupted altogether (especially if the parts are worn from extended use).
Many of the problems of BNC connectors can be traced to aspects in the design of the male half, which is to say, the part that has the male center conductor pin and that is given the quarter turn twist while gripping a knurled shell we shall call a bayonet latch. Let us briefly take a closer look at the conventional BNC connector, the better to appreciate why it has these problems.
The female connector portion includes a female center pin that is centered and held in place by an enclosing Teflon female sleeve. The female sleeve has a reduced diameter portion in front, and toward the rear has a stepped diameter that engages a corresponding shoulder in a female shell. The female sleeve is secured in place from the rear in various ways, depending upon the style and manufacturer. The reduced diameter portion in front will be of interest, shortly.
Now consider the male connector half. As an assembly, it includes a Teflon male sleeve whose rear portion has a small diameter bore that centers and supports a male center pin, and whose front portion has a larger diameter bore sized to just slip over the reduced diameter portion of the female sleeve. When the connector halves are properly mated the two Teflon sleeves are not only in contact over adjacent cylindrical surfaces, but the female sleeve xe2x80x9cbottoms outxe2x80x9d inside the male sleeve. (The terms xe2x80x9cmalexe2x80x9d and xe2x80x9cfemalexe2x80x9d are applied to component parts according to the connector halves as a whole, and its gender is determined by the shape of the center conductor pin. Viewed in isolation, the xe2x80x9cmalexe2x80x9d Teflon sleeve might be thought to be xe2x80x9cfemalexe2x80x9d, as it surrounds the outside of the xe2x80x9cfemalexe2x80x9d sleeve when the connector halves are mated. But it is part of the male connector half. So it is that the male sleeve has a female shape, but is still called the male sleeve.) Potential gender confusion aside, the important thing is that when proper mating occurs there are edges and surfaces of the sleeves that xe2x80x9cvanishxe2x80x9d to form one (i.e., unitary) longer tube of Teflon that will be the dielectric material disposed between the center conductor and the outer shield forming the coaxial transmission line.
A similar thing happens to the center pins that they carry. The male pin has a reduced diameter tapered tip that enters a cavity, or socket, centered in the end of the female center conductor. The cavity is slightly undersize, but the end of the female socket is slit to allow a slight resilient outward motion that promotes good ohmic contact between the pins. The thus-expanded outer diameter of the female center pin is the same as that of the male center pin, so that when they are fully mated a shoulder on the male pin and the face of the female pin xe2x80x9cdisappearxe2x80x9d as each of the two pins presses against the end of the other, and the pins appear to be one (unitary) longer cylindrical conductor.
The two sleeves and the two pins are supposed to fully mate simultaneously, for if one were to mate before the other it would prevent the further motion needed by the other to become fully mated.
Surrounding and carrying the sleeves are respective cylindrical connector shells, one male and one female. The male shell has a collection of slits so that they can bend inward slightly under compression as they enter a female shell of slightly insufficient diameter. This provides good ohmic contact for maintaining the outer shield of the coaxial system. Once again, the male shell is expected to bottom out against a stepped diameter within the female shell, so that (save for the slits) the mated pair of shells appears as a complete unitary cylinder of uniform inner diameter as the end of the male shell vanishes against the shoulder inside the female shell.
A pair of bayonet pins on the outside of the female shell engage detents at the end of a quarter turn spiral groove in a rotatable captive bayonet latch carried on the male shell. Depending upon the particular design, a spring located somewhere in the above described elements provides a resilient force that pulls the center pins, sleeves and shells together once the detents in the bayonet latch contain the bayonet pins. If everything is working correctly, no RF currents flow through the connection between the bayonet pins and the bayonet latch. Unfortunately, pulling on the cable, or otherwise inducing external tension urging the two connector halves apart, can overcome the internal spring tension keeping the connectors halves together. If a sufficient tension is applied the connector halves will draw apart slightly.
There are two basic aspects that we wish to point out. First, the tapered end of the male center pin enters a slitted socket in the end of female pin, and ordinarily spreads those slit portions apart slightly, for good contact. As the connector wears the diameter of the tapered end portion of the male center pin and the resilience in the slit female pin are both reduced, while the inner diameter of the female pin is increased, so that a slight withdrawal of the male pin can significantly decrease the ohmic quality of the connection. Equally as bad at higher frequencies, as the withdrawal occurs, there appears a short length over which there is a marked decrease in center pin diameter. That is, the male and female center pins have the same outer diameter, and when they are fully mated there are annular surfaces that touch, shoulder to shoulder. When that occurs there is no, or very little, effective change in the outer diameter of the combined center pins. When these shoulders do not touch there is an immediate reduction in diameter to that of the tapering end of the male pin. A similar increase in the effective diameter of the outer conductor occurs also, as the end of the male shell pulls away from the shoulder in the female shell that it seats upon. These changes are important, since the characteristic impedance of a coaxial transmission line involves the ratio of the outer diameter of the center conductor and the inner diameter of the outer conductor, as moderated by the dielectric constant therebetween. When the male center pin withdraws slightly from the female pin, the short length of diameter reduction occurs at about one quarter of an inch from the location where the short length of outer diameter increase occurs, and this xe2x80x9cdouble whammyxe2x80x9d appears as a very definite discontinuity.
A similar bad thing happens in connection with the Teflon sleeves. Ordinarily, the reduced diameter section of the female sleeve would be the exact complement of the large diameter portion of the male sleeve. The idea is that when they mate their edges vanish, as it were, and the two parts act as a single part of continuously present material of the proper diameter. That fails when the connector halves pull apart, producing another discontinuity owing to a location of altered dielectric constant. This happens adjacent where the center pins have their xe2x80x9cdiameter fault,xe2x80x9d increasing the resulting discontinuity. Furthermore, the presence of the Teflon is a bit of a problem in the first place, since it is difficult to machine the stuff to the tolerances needed to reliably perform the magic of the vanishing edges. Also, it is the Teflon that is supposed to hold the center conductor pins in their proper locations. Not only is Teflon difficult to machine to tight tolerances, but it won""t hold them over time, even if it could be done, since Teflon cold flows so easily. Even a brand new connector, but especially a used connector, will have Teflon sleeves that exhibit and account for significant mating anomalies at frequencies above, say, 500 MHZ. This is no longer a minor matter.
Here now is a brief summary of how the improved BNC connector described in PRECISION BNC CONNECTOR solves these problems. Here is the Summary Of The Invention from that Disclosure:
xe2x80x9cA solution to the problem of poor RF performance in the conventional BNC connector is to first, eliminate the use of Teflon, in favor of an air dielectric in the vicinity of the mating parts, and support the male and female center pins further back within the body of the connector, using other proven dielectric materials borrowed from the precision type N connector, or from another 7 mm RF connector. Next, a captive knurled draw nut provides positive displacement and the tension needed to draw the already mated male and female connector halves together, in place of the conventional spring tension. It is the bottoming out of the male shell inside the female shell that resists the positive displacement and the tension supplied by the knurled draw nut, ensuring that the two connector halves are actually in contact, and that the edges of surfaces that need to xe2x80x9cvanishxe2x80x9d for good operation do indeed vanish. The mating center conductors are rigidly mounted within their shells and bottom out against each other at the same time as do the shells. The basic bayonet latch mechanism is retained, so that either half of the new connector will mate with opposite sex halves of conventional BNC connectors.xe2x80x9d
Today, many oscilloscopes operate at ten times the frequency at which conventional BNC connectors begin to exhibit degraded performance, and some operate considerably higher. There is, in fact, a large installed base of such oscilloscopes that use a conventional BNC connector. These high frequency ""scopes use active probes that perform, among other things, impedance conversion, so that the signal can be supplied to the ""scope over an intervening 50 xcexa9 transmission line, which is the cable that connects the probe to the ""scope. We are now faced with a situation where the connector of choice is a principal limitation in the overall performance of the ""scope/probe combination. It is true that there are other RF connectors that would solve the problem of the rotten RF connection, but they are unsuitable for one or more reasons. Some are simply too expensive, and, it will be noted, the expensive ones tend to be threaded and/or easily damaged; APC 3.5 connectors come to mind in this regard. Precision type N connectors would carry the signals all right, but they, too, are threaded, and besides being moderately expensive, they take a lot of panel space. The old GR-874 xe2x80x9csexlessxe2x80x9d and xe2x80x9cpush-onxe2x80x9d connectors even comes to mind. It was (and still is!) a pretty good connector, perhaps when in good condition, even comparable to precision type N. But it is as big or bigger than N, is more expensive, and sadly, seems to be on the verge of xe2x80x9cgoing away.xe2x80x9d Well, then, so be it. It would seem that we should switch to the precision BNC connector. (We note that it cooperates, with some degradation in performance, conventional BNC. That helps lessen the sting of a change to a new style.) We can easily arrange to use the precision female portion on the front panel, since it is essentially a direct replacement. Alas, even if we do, there is yet another fly in the ointment.
The second issue concerns the electrical attachment of ""scope probes in particular. In the oldest (and by today""s standard, largest) passive probes, adjustable compensation was located in the probe body and the cable at the ""scope end had just a boot protecting a cable mounted male connector. Front panels were big, bandwidth was low, and this was thought to be a tidy solution. Later, for smaller passive probes of higher bandwidth the compensation components were located in a small box at the ""scope end of the cable, and a bulkhead mount male connector attached the box to the female bulkhead connector on the (smaller) front panel of the ""scope. Today""s very high bandwidth active probe is smaller still, and for some brands the ""scope end of the cable has a pod or housing the size of a small farm-rat (or at least a large house mouse) that contains a xe2x80x9cpush-lockxe2x80x9d BNC connector, and also provides mounting for a modest number (six to nine) of other single conductor auxiliary connections between the pod and the front panel. There are many reasons to have this housing in the first place, and good ones for having it abut the front panel of the ""scope. Probe identification, probe settings, probe power (and possibly, but not necessarily, power return) are all conveyed by these additional connectors (which are essentially spring loaded pins). The push-lock feature arises from the need to do something to cause the quarter-turn twist that the bayonet locking mechanism requires on the one hand, and the desire to not require rotation of the housing, lest that cause mischief from temporary mis-connection between the pins and their corresponding pads on the front panel. Add to that the circumstance that there is (as a practical matter) no room to get a user""s thumb and forefinger in there to rotate the BNC latch. For one thing, the face of the pod or housing should be up against the scope front panel to assist in making the auxiliary connections, while for another, it is sometimes the case that adjacent BNC jacks are located so close together on the panel that, even if there were no rat-sized bulge in the way (and perhaps no auxiliary conductors), it would still be a real aggravation to get that thumb and forefinger in there to twist the BNC latch.
The push-lock BNC connector described in the incorporated ""841 Patent solves this issue nicely. One merely holds the pod or housing in the hand, and while the connector halves are axially aligned, pushes the housing toward the ""scope. The assembly in the housing that corresponds to the BNC latch twists, but not the pod (which may even have alignment tabs to prevent it). Eventually the detents of the twisting latch align with the bayonet pins of the female connector half on the front panel, and spring bias rotates the latch by an amount sufficient to achieve engaged detents, or xe2x80x9clockxe2x80x9d (which is less than the usual quarter turn). To release the male connector/pod the user presses against and rotates a tab or lever with his thumb or a fingertip. The tab is a portion of the BNC latch mechanism that extends out from the pod or housing for just that purpose. Once the latch is rotated to clear the detent, the user simply pulls back on the housing to separate it from the front panel. Unfortunately, despite its ease of use in attaching and detaching it from the ""scopes front panel, it is still a conventional BNC connector as far as the quality of the transmission line segment formed by the connector is concerned. It significantly limits the performance of the ""scope when higher frequencies are considered.
It will be clear that what we ought to do is put a male precision BNC connector into the housing, with the intent of having it work in a manner similar to the way the push-lock technique presently does. But upon reflection we realize that, once the detents in the BNC latch engage the bayonet pins, the precision BNC connector utilizes a knurled draw nut that engages threads on the male side and then draws the two connector shells tightly together by retracting the male BNC latch against the female""s bayonet pins, thus forcing the shells and center pins into complete contact. (Another way to state it is that tightening the draw nut forces the male shell and pin forward into their female counterparts, and after they simultaneously bottom out, continued displacement pulls the BNC latch away from the bayonet pins. That provides the static force that keeps things tight.) To have to additionally turn (and un-turn!) that knurled draw nut seems like a form of retrograde progress, compared to the ease of use presently associated with the conventional push-lock connector technique described in the incorporated xe2x80x2841 Patent. But it has the potential for 18 GHz performance when both sides are precision BNC, and it even still mates with conventional BNC connectors on existing xe2x80x98scopes (although with same reduction in the degree of increased performance, owing to the presence of Teflon, tolerances, etc.). This is just too good to pass up. There has to be a way . . . the question is: xe2x80x9cHow to do it?xe2x80x9d
A housing carries a male push-lock precision BNC connector. The function of the BNC latch of the male connector carried therein is performed by two cooperating parts: The first is a spring-biased lock ring captive in a housing and that fits snugly and concentrically over a male sleeve. The second is the male sleeve. There are different embodiments, but in one an end of the male sleeve is pressed over a back portion of a BNC male shell, resulting in two rigidly attached cylindrical portions that are coaxial, coextensive, and separated by a slight gap wide enough and deep enough to accept a BNC female shell during connector mating. For convenience, we call this subassembly a double shell. The double shell is held captive, against any motion, by anchoring it to the inside of the housing. Thus the inside cylindrical portion of the double shell functions as the BNC male shell required for BNC connections. The outer cylindrical portion (the male sleeve) has a slot and/or cut-away portion to accept the forward travel of the bayonet pins, which, when the connector halves are engaging, extend beyond the thickness of the sleeve and into the region occupied by the lock ring. The lock ring has grooves having an angled portion that engage the bayonet pins of a female BNC connector. As the housing containing these male parts is held without rotation and moved toward the female connector, the lock ring rotates as the grooves contact the bayonet pins (which don""t move, either). After sufficient angular rotation the angled portion of the grooves (as seen by the bayonet pins) bend to run parallel to the axis of the center pins, allowing the two connector halves to approach each other without further rotation of the lock ring. After the male and female halves are essentially fully mated along the axial direction, the axial portions of the grooves clear the bayonet pins, and a bias spring rotates the lock ring, whose grooves now present a path at right angles to the axial path. Rotation of the lock ring now blocks the separation of the connector halves, by interfering with where the bayonet pins would attempt to travel if the connector halves were pulled apart. Such pulling apart forces the front of the lock ring against a surface on the inside of the housing, where that motion is resisted. The result is that the BNC latch is captive on the bayonet pins.
To this point we have described the push-lock mechanism, equivalent to what is set out in the incorporated ""841 patent. To create the tension needed to draw the precision BNC connector parts firmly together, the grooves resume their angled path after a brief length for the right-angled region. The new angle is shallow, and will be called a ramp portion. The bias spring continues to provide force to rotate the lock ring, and eventually the bayonet pins resist any further motion of the lock ring. The lock ring has a thumb tab that extends out from the housing to allow the motion of the lock ring to be reversed when removing the push-lock connector. During installation the operator may, if desired, use his thumb or a finger to push the other way on the tab to assist the bias spring and further wedge the shallow ramp of the grooves tightly under the bayonet pins, where in any case, they stay until released by the normal release operation just described. The ramp portion is shallow enough (say 12.5xc2x0), and the bias spring force sufficient, to reliably wedge the bayonet pins until so released. Wiggling the housing actually allows the bias spring to increase the wedging action. The resulting force between the bayonet pins and the grooves in the lock ring performs the same function as the knurled draw nut in a stand-alone precision BNC connector.
In principle, the male sleeve may have only slots to accept the travel of the bayonet pins. In that case, however the amount of penetration of the bayonet pins into the grooves of the lock ring is diminished by the thickness of the sleeve, and may not be enough to assure reliable operation. The slots can be widened into cut-out regions proximate where the grooves in the lock ring are, and then the lock ring simply made thicker in those places, so that it descends into where the sleeve is now cut away, as it were. This allows the grooves to be deeper and contact the bayonet pins as effectively-as a conventional male bayonet latch does. There are yet other variations on how the sleeve is effected, some described in the incorporated ""841 patent.
It will further be appreciated that the lock ring described for a precision BNC male connector, when used in an otherwise conventional push-lock connector having standard BNC male parts and mated to a conventional female BNC connector (as in an older ""scope that is part of an installed base) will also exhibit improved performance, owing to the BNC shells being drawn together and minimizing the xe2x80x9cnormalxe2x80x9d discontinuities. And, of course, a new push-lock precision BNC (male) housing would also mate with an older female BNC connector, again with partially improved performance.